Abstract
Motivated by different psychophysical results it has been proposed that perception and execution of actions might share a common representation (Hommel, et al., BBS 2001, 24,849). Recent neurophysiological and fMRI experiments showing that some cortical areas are active during both, action observation and action execution provide support for this hypothesis. One prediction of such “common coding” theories is that internal representations acquired through the learning of new motor skills should influence the visual perception of similar motor behaviors. We evaluated this prediction by testing whether motor learning, in the complete absence of visual feedback, can influence the perception of biological motion. A set of point-light walkers was created by changing the relative timing of the limbs. While maintaining the synchrony between contralateral arms and legs, the relative phase between the arms was varied resulting in frequent gait patterns (relative phase close to 180 deg), and very infrequent gait patterns (relative phase strongly deviating from 180 deg). Subjects were tested in a same/different psychophysical paradigm using these stimuli. After this test they were blindfolded and trained to execute one of the infrequent phase relationships with their arms. During this training only non-visual feedback was provided. Also any cues that would provide explicit temporal rhythms were avoided. The psychophysical test was repeated after this motor training. Our preliminary data show that before training performance of subjects in the “same” trials was higher for frequent than for infrequent phase relationships. After successful motor training, performance increases selectively only for the trained infrequent phase difference, even though no visual input was provided during training. Our results seem consistent with a top-down influence of internal motor models, or abstract representations for relative timing on the recognition of biological motion.
Supported by the Deutsche Volkswagenstiftung and the Hermann and Lilly Schilling Foundation. Additional support by the MPI for Biological ybernetics is gratefully acknowledged.